Train motor sound - What UK enthusiasts get wrong while Japanese enthusiasts build model VF drives

[@AegisEdge]

as a Japanese person who is a big Japanese rail enthusiast, thank God someone pointed this out. Whenever I see the brits write a whole essay about how VVVF's work and the difference IGBT and GTO, or where the sounds come from. They just don't have a clue and it drives me up the wall lol.

[@dennismundt7378]

An interesting video. The European locomotive Siemens ES64U2 also plays a musical scale when accelerating and braking. Currently, the powerheads of the German ICE-1 trains have also been converted from GTO converters to IGBT converters for future use. This also changed the acoustic frequency of the traction motors after the conversion to IGBT converters when the train started up and accelerated.

[@no-damn-alias]

They also did that to the SBB Re460

[@ZK5120A1]

GTOs and IGBTs are just switches. The sound is defined by how these switches are controlled to turn on and off (the inverter program) but not the switches themselves IGBTs are able to be switched at a higher freq, but if you try to switch them with a lower freq, the sound will be just like an GTO inverter

[@ArneChristianRosenfeldt]

..And the actual switching edge is so sharp in both devices that any audio effects from it are far beyond the hearing range of a human.

[@gorak9000]

I have come from the future from just before the universe ends and on behalf of the entire internet to award you with the best use of MS Paint in the entire universe and of all time to explain a complicated subject award! Congratulations!

[@SP_1900]

Thanks for clearing up the confusing explanations that I’ve seen on many articles for UK trains. Now it is easier to figure out how to program the sounds in

[@drawingboard82]

Excellent explanation, many thanks, drives me nutty as well. As you stated, the sound itself comes from the changes in switching frequency. This switching frequency vibrates the cables, housings and motors, making them act like a loudspeaker. A couple of other points: 1: as well as speed control by varying the output sine wave frequency, most vf controllers for AC motors have torque limiting features. On an induction motor the torque is proportional to thr current and the current is proportional to the slip. The slip is the difference between the rotor speed and the soeed of the rotating magnetic field, usually around 5%. By controlling the output ac frequency to limit slip, current and therefore torque can be controlled or at least limited. 2: everything in this video refers only to trains with AC induction motors with VF multi phase drives. These drives are pretty standard on all electric trains from the mid '90s onwards. As far as i know they are all three phase although in principle more phases could be used. The British type 45 destroyers have ac induction motors and can operate on 3, 9 or 27 phases, the higher phases being less efficient but quieter. 3: older trains with dc motors or (mostly in Germany and other 15kv, 16.6 hz supply countries) ac series wound motors will make a variety of sounds. The more modern dc traction trains (mid '70s to early '90s) used thyristor choppers but because the dc motors dont require an ac waveform the thyristors don't need to change switching frequency and the hum is constant. 4: The newest trains such as the ice 4 and many new trams are using permanent magnet brushless synchronous motors. These have a slightly higher efficiency and, usually, lower space and weight requirements. They usually require closed loop control with feedback to operate efficiently and limit torque. I don't think these change switching frequency as much but it probably varies by manufacturer.

[@petemoss1938]

Thanks BENO, this will help me explain to my son, an aircon engineer how VF drives work. Used more and more in the HVAC industry!

[@graythewolf6096]

I would like to note at 2:20 in some cases with 3-level inverters the power could be at 50%. This is used in cases such as the JR East E231-500 previously used on the Yamanote Line. Many Shinkansen bullet train models also use this inverter circuitry including a few other variations of the E231 and other classes within the Japan Railways fleet. The DB BR 481 also has this capability within it's inverter circuitry from what I understand, however this is never used and operates in a standard 2-level PWM mode. Thyristor and Transistor switches can not be 50% on their own however, and require some form of resistance to create the 3-level effects.

[@benolifts]

I would assume that there are multiple thyristors to create this, rather than a thyristor being run at 50%

[@coastersaga]

What about the L0 series maglev that will be used on the Chuo line? Does it have a VVVF drive?

[@calvinkulit]

​​@@coastersagaI believe that just uses rapidly switching electromagnets for propulsion

[@nooxis2643]

@@coastersaga afaik the latest technology used on the test line was IEGT cycloconverter

[@Lucius_Chiaraviglio]

@@benolifts Presumably 1 thyristor/transistor for generating voltage >0 and 1 thyristor/transistor for generating voltage

[@AG-un7dz]

This deserves two thumbs up!

[@LiftSurferRYEGI]

A very well explained video. I used to know a train enthusiast that is pretty dumb too.

[@FAB1150]

For anyone confused by how the chopping actually translates into sound: Those drives work by pulsing the voltage so that the average voltage during that period of time is what the voltage of a real sine wave be at that time. But they still "chop" the voltage, alternating from high to low. When voltage is high, current starts flowing into the motor. When current flows, the motor pushes the cart, when no current flows the motor doesn't push. When the vfd chops the voltage fast enough, the capacitance and inductance of the motor, wires and circuits* actually smooth out the chops making what's close to a real sine wave (basically "calculating the average" in a very analog way), but when it's low you get the motor actually alternating push/not push at the frequency of the chop. The pulsing motor creates vibrations and voilà! You get the noise. Now you can understand the rest of the video better :D *What does this mean? Well, a simple explanation is that the motor windings need a bit of time to "charge" and produce the magnetic field to push/pull the magnets, so the current doesn't immediately go to "max", and then to come down to 0 (that's called indictance!). Think of it like bagpipes! You blow into the bag quickly and then take a breath while the bag is "charged". If you alternate breathing and blowing fast enough, you can make a steady sound! If you take too long of a breath you run out of air and... You can imagine wgat happens :D

[@benolifts]

Yes, but there are parts you could have said clearer. The number of pulses per second acts as a speaker, which is what makes the sound, while it is the frequency of the sine wave, made from the pulses, which gives the output frequency. Your description kind of makes the distinction between the pulsing frequency and the output frequency not as clear as it could be.

[@FAB1150]

@benolifts yes, that's what you explained well in the videos! I was strictly speaking about how the motors actually "translate" the pulses into noise. Something a bit more niche but that somebody might be curious about. Cheers!

[@ArneChristianRosenfeldt]

So is this like a piezo speaker ( PC speaker, cheap alarm clock ) or do we hear the sine wave imperfections in the magnetic field ( Lorentz force or magneto-Striktion ?).

[@FAB1150]

@@ArneChristianRosenfeldt it doesn't act like a piedi speaker, but like an "actual" speaker, the classic ones. Speakers are actually exactly the same, but linear! They're optimized to make noise, but in the end they're linear motors. Here, it's doing the same thing but while rotating! What you hear is the shaft quickly alternating between pushing on all the mechanical components of the train (when the pulse is high or low), and not doing it (when the pulse is at 0).

[@ArneChristianRosenfeldt]

@@FAB1150 To state it more clearly: "close to a sine wave" for humans sounds like a sine wave + a chirp. I think that this is a numbers game. Trains are so powerful that even a slight deviation from sine is audible. Kind like cars are so loud. I mean, automotive cars. With fast enough chopping in small motors at each axle, the noise from the track becomes louder at 30 km/h. Track noise is vertical, while torque noise moves the wheel forward and backwards -- haha.

[@amiausUSA]

While visiting Austria a few years ago, I heard similar acceleration noises from the ÖBB locomotives. Those were mostly from the Railjet trains.

[@UserCraft1]

yes, we use the siemens taurus locomotive for the railjet, and it has a distinctive startup sound, "the tone ladder". I've noticed it when standing by a departing trains, and it seems to stay at a fixed sound after a few seconds.

[@andrewthompsonuk1]

Great information, I had always known that the noise was due to the switching strategy but not what the strategy was.

[@Lucius_Chiaraviglio]

I'm in the same boat. Although I thought they might be also doing something like switching the number of motor poles to keep the switching frequency within acceptable parameters for the transistors/thyristors.

[@hotmixer2010]

There's a new type of VF drive used in trains SIC-VVVF instead of IGBT, seems like the new VF drives are smaller and more compact

[@RockstarRomania]

Technically more correct would be MOSFET-VVVF, as that's what's directly replacing the IGBTs.

[@gorak9000]

I"ve heard some rumblings in the last few years about silicon-carbide power electronics - it wouldn't surprise me if they're making a debut in the mass market by now.

[@ThompYT]

​@@gorak9000they are in some electric cars like teslas

[@Lodai974]

@@gorak9000 The SiC VVVF is already on the new Japanese trains...and also for example Bombardier (Alstom), the SiC Mitrac, audible in Singapore and Stockholm.

[@Pippi444]

I had a dream where i was on a vintage subway train that had an epic sounding loud VF-drive that sounded similar to the first japanese VF-Drive you showed but instead of musical scales it played a few notes from Justin Bieber's song "Baby"

[@iancastleton9052]

Brilliant explanation! Thank you for this.

[@sgbuses]

I am inclined to believe this explanation much more. For many years I really struggled find an exact explanation on why Singapore's Alstom Metropolis C751A, which on specification says repeatedly stated that it runs on Alstom Onix IGBT-VVVF, sound nearly identical the London 1996TS which run on GTO-VVVF. Any other explanation offered previously would logically implied that the C751A is GTO-VVVF, which it isn't.

[@crazyleyland5106]

I like the sound that 1959 stock made. As a much older electric technology. I assume that tube trains newer than 1996 stock don't have pulsing motors, but I haven't been to London for years. Nb a new genre of electrical videos has come out recently from Japan and South Korea - overvolting toys!

[@benolifts]

All modern trains, electric cars, electric buses, E scooters, and pretty much all traction motors use the method described in this video. Although in smaller motors the carrier frequency would be so high that it never reaches the stage where it has to switch to pattern mode.

[@coastersaga]

@@benolifts What about using a pure sine wave oscillator?

[@benolifts]

@@coastersaga It is not possible. Think of ohms law. Any part of a circuit that is resisting current is dissipating power. Power can not simply be resisted or turned down like a volume control. The motors are large. A lot of current is passing through them. Large amounts of current can not have the flow of current impeded without huge amounts of heat being produced. The only way current could be resisted is by having motors in pairs and switching between series and parallel, or by using huge heat dissipation resistors, which is what old DC trains used. But with VF drives and AC motors, the output frequency has to be changed. The only possible way is by PWM, so the motor is switched on and off at a very fast speed to create an average voltage.

[@bunnymaid]

I think they retired the last of the Keikyu 2100 VVVF trains recently. The cute thing is, they were originally GTO, but the replacement controllers were IGBT.

[@AegisEdge]

no it wasnt even recently it was ages ago. They more recent were the N1000 series

[@robertcrossland2754]

Finally, a brilliant explanation!

[@pieterpretorius1014]

i always thought it was the coils whining from the current going through the coils at specific frequencies. i have observed similar whining from stepper motors and rc brushless motors but just more constant and not the stepped whines from train motors

[@officialmcdeath]

Damn, these Japanese make me want to go back to Austria and those wonderdful 1016 and 1116 Taurus locos \m/

[@lzh4950]

The Keikyu 1000 series rolling stock used to have propulsion systems from Siemens too I remember (the one in this video)

[@FerrybigGaming]

2:33 Only class AB radio's are doing this, the more common class D radios use PWM

[@BenDeSwert666]

Nice explanation. I'm impressed. 👍🏻

[@EE12CSVT]

Many thanks for that, it's explained it very nicely.

[@Lucius_Chiaraviglio]

Is the limiting factor on transistor/thyristor switching frequency due to the time to go from 0% to 100% or 100% to 0% actually being non-zero and NOT possible to reduce proportional to the desired switching frequency, resulting in the transistors/thyristors actually spending some time in a partially resistive state, and the proportion increasing as the switching frequency increases, thereby causing them to overheat if switched too fast? I would guess that switching too rapidly would also cause eddy currents in the windings and associated materials interacting with the magnetic field, and these would also generate heat without doing useful work, like in a transformer that is fed with too high a frequency (for which its lamination is not fine enough) -- although since as shown in the oscilloscope trace, the current to the motor is partly smoothed out, it's probably the overheating of the power electronics that will usually get you first.

[@benolifts]

Your first paragraph is 100% correct. It is the time spent on when neither 0% or 100% which is the issue, so as the motor speeds up the pattern as to be reformed into simpler patterns at set intervals to reduce the number of switching states. With your second paragraph, its actually the opposite that's true. The faster the switching frequency that more accurate the sine wave, so the non desirable currents are in the motor. A lot of smaller motors in industry nowadays have the VF drive switching set to 10,000hz.

[@Lucius_Chiaraviglio]

@@benolifts So in other words for the second part, the greater accuracy of the sine wave more than makes up for the tendency of the higher frequency to make eddy currents?

[@Lodai974]

To put it very simply, the audible sound comes from the stator of the motors which react to the received switching frequency. Most traction systems using inverters to manage current inputs to the motors use modulation starting with a PWM (asynchronous), then synchronous cutting schemes (SHE 7, SHE 11 etc.)

[@bb-3653]

I remember seeing those japanese videos. They are so cool

[@brlinf06398]

2:17 looking at the image, so why don't we also do this with speakers

[@BrickwallCollision]

"class d amplifier" 👍

[@meki___6881]

this video gives me a crazy idea. are electric motors silent?( ignoring bearing and brush noises) and all electric motor noise comes from phase shifting both pvm signals or rotary phase shifting in a brushed dc motors?

[@benolifts]

Electric motors make 3 sounds... -The PWM pulsing (the motor acts as a speaker caused by the pulsing in frequencies within the range of human hearing) -The motor bearing sound (low humming sound when near full speed) -The sound of the current in the electric motor (the traditional sound of a motor including in the past before VF drives, this sound is louder the more power being put into the motor, and the pitch is the motors speed) In addition to this on a train you can also hear the sound of the wheels on the track, also there is the 50/100hz sound of the rectifier

[@Mladjasmilic]

So, it is possible to make an electric car, that has shifting in VF drive? And it could be made in hearing range frequencies. For those who straight-pipe their exhaust, they could remove the noise filter to make it louder.

[@benolifts]

Toyota Prius does this. You have to listen carefully as it is doing it at quite a high frequency. 0-20mph is at pure PWM, then 20 to 45mph is using 3 different pattern based PWM. The motor has a top speed of 45 mph, then the power split transmission's CVT system diverts more power to the engine and the motor's speed is held by the CVT to stop it over speeding.

[@HansTheGeek]

Is the motor or the electronics making the sound?

[@benolifts]

The motor acts as a speaker and makes the sound caused my the switching.

[@no-damn-alias]

As this on the topic of converterts. Do you repair ac/dc converters or do you know anybody who does

[@theblockybanana5537]

The first VF drive simulator you showed had a changing carrier frequency, it increased. Why did it do this?

[@olhoTron]

I am by no means a expert, but I think it's for eficiency reasons, a higher carrier frequency means more switching losses And on acceleration from a stand still is when the train uses the most energy, so efficiency is king

[@theblockybanana5537]

The transistors are turned on and off to produce an AC sine wave. But what keeps the transistors working in time? What tells them to turn off and on? How are their timings controlled?

[@sagnost]

Microcontrollers?

[@theblockybanana5537]

@@sagnost and what controls their timings?

[@sgkingly8392]

@@theblockybanana5537 I'm a mech engineer not electrical but have some understanding of control systems. Control electronics have a clock which is a resonator that resonates at a constant frequency. Operations are based off of that clock and its what governs the time. When operating in PWM mode the whole thing has a duty cycle that is some fraction of the clock frequency. A duty cycle of 100% means that the signal is on all of the time and a duty cycle of 0% means that it is off all of the time. There will be a reference signal (Motor rpm, output voltage or specified waveform etc.) and output sensors measure the output of the controller and compare to the reference. The duty cycle is then adjusted accordingly to maintain that. When operating in pattern mode the same clock is used and pulse widths remain constant but the frequency is adjusted to maintain the desired output again the clock is keeping time. This is also a non dynamic method where a set frequency of pulses is used for a set motor rpm rather than using sensors and a closed feedback loop. Take all that with a pinch of salt though, I have taken a few control system classes and done some PWM programming but my knowledge is very surface level, it is not one of my good areas

[@drawingboard82]

The inverter control computer is in charge of switching the semiconductors (igbt or gto) on and off. How it does so is usually software configurable. Almost all VF drives for AC induction motors estimate motor speed via back emf and control (or at least limit) torque by matching the output sine wave frequency to the rotor speed plus a maximum.

[@liujiaming2710]

what I didn't get is, that when the pulse is cleaned say from 10 pulses to 5 pulses per cycle, why does not the output frequency drop? Say the output freq can stay at 10Hz when the pulse is lowered to 5 per cycle and increase as pulses increase again, why didn't it just use 5 pulses to output that 10Hz in the first place?

[@benolifts]

The issue is a thyristor must never resist current. A thyristor must only act as a switch. It can not be a resistor. Heat dissipates at the point of resistance (ohms law) so the thyristor must do a clean switch between off and on. However, as it is not a perfect world, the change from on to off is not quite instant, there will be a slight moment (just a nano second) when it isn't a clean switch and is between on and off states, at this point heat is dissipated. On pattern mode as the output frequency speeds up the switching also speeds up with the pattern. This means that the thyristor changes state more and more often. Due to this the number of nanoseconds where the thyristor is neither on or off increases. This is the reason why the pattern must reform with less pluses. The reduction of pulses when changing from, lets say, 7 per half cycle to 5 per half cycle means less accuracy on the artificial sine wave. However the frequency of the sine wave is not affected.

[@Andrewjg_89]

Older electric trains have better traction sound including the Class 323, Class 465 which have been modified with Hitachi traction motors and of course the Jubilee Line 1996 Stock built by Alstom. Even the newer ones also have epic traction sound like the Class 357, Class 375, Class 376 and Class 377 Electrostars built by Bombardier.

[@JackHymers-cq6kt]

Some 465s still have their original motors

[@windowsxpnt2347]

Hey Austin this is guys

[@ReubenAshwell]

Hey guys, beware of misinformation.

[@LiftSurferRYEGI]

Hey guys this is surfy

[@lo2740]

the Japanese are just another level, another league.

[@clementpoon120]

the japanese motor controllers are like the nerdiest thing that ive ever seen, i love it

[@youkofoxy]

I am Having a blast with this 100% unadulterated autistic model train power train rant.

[@mariog1051]

What language is the speaker using?

[@benolifts]

How do you mean?

[@mariog1051]

@@benolifts sorry, I meant accent, not language

[@gabrielstravels]

​@@mariog1051He's from the UK, so he's British

[@benolifts]

@@mariog1051 I would describe my accent as a mix of British RP English, British Chav English and British Autistic English.

[@Plaxton_presidents]

You would like the plaxton president I rode the other day the VF was very loud

[@benolifts]

Plaxton president doesn't have a VF drive. Well maybe the fuel pump could possible be on some sort of drive controlled motor, but you wouldn't be able to hear that.

[@Plaxton_presidents]

@@benolifts I have a video of it somehow it’s on my channel

[@Plaxton_presidents]

@@benoliftsI found out what it was it was probably the gearbox

[@S2artrail]

Anyone ready to tag Armstrong Powerhouse?

[@RSX798LockedandLoaded]

Inyong Napakinggan ang Kabuuan ng GMA Super Radyo DZBB... *SUPER BALITA!!!*

[@phoenixarian8513]

You probably make it wrong. PWM mode (non-sync mode in japan) does NOT need to output 50Hz when your train is starting! This can't be more obvious since it's starting the 3-phase AC motor wants some REALLY LOW frequencies! It can be as low as 5 Hz so you can have 80 ticks in each cycle (or 40 ticks per half cycle) to switch between on and off. And 40 cycles can contain a 5-pulse mode or 3-pulse mode if you carefully pick the order to switch. Still the tickrate resources are pretty scarce in this scenario therefore switching to pattern mode (sync mode in japan) is the better idea.